next up previous contents
Next: A closer look Up: Flash 92 Previous: Observations of Molniya

Determination of Rotation Axis Project

The experiments with the recently developed software for determination of the rotation axis of an end-over-end tumbling satellite have been rather successful, as reported elsewhere in this issue (and Flash 90 and 91).

I would like to start a project aimed at determining the rotation axis of several flashing satellites. The project would involve quasi-simultaneous timings of as many flashes as possible by observers located at different sites.

Using these timings it should be possible to determine the rotation axis of a flashing satellite rather accurately. The experiments in the last two months have shown promise (see my message on the Sampex rocket in Flash 91 and also the article in Flash 90).

It is now time to start a regular observing program to track the evolution of the rotation axis of several tumbling satellites.

Simultaneous observations of a tumbling rocket are valuable, that's quite clear, but very difficult to do. It is not clear to me at this point that the difficulties on the practical side are compensated for by higher accuracy of the determined rotation period and axis.

Quasi-simultaneous observations are also very valuable, and much easier to do. Basically, it means everyone tries to make as many observations of a small group of specific satellites as possible. Uncoordinated, yes, since for quasi-simultaneous observations we only need observations within a time range of a few days (24 to 48 hours).

This was demonstrated in Flash 90 for the satellite 82- 40 J, for which we had three short series of timings by two observers over a range of more than 24 hours. The basic assumption behind the validity of 'quasi-simultaneous observations' is that the direction of the rotation axis does not change much over 24 hours, something which we think we can justify from our understanding of the theory of eddy current deceleration of tumbling rockets, see the Brochure. Another assumption is that the rotation period does not change much over the same time range. Both assumptions are OK (we think) for decelerating rockets, and probably also for accelerating objects (at least within the bounds of our accuracy).

Multiple Timings

The goal is to accurately measure the time of all (primary) flashes during the pass of one of the satellites in the list (further on in this text). There are basically two techniques for this.

The first involves using a stopwatch that can store 50 or more split timings. Each time you see a flash, you push the button. At the end you stop your stopwatch on a known time signal (radio, GPS receiver, etc...) so that the time measurements are absolute. Afterwards you write down the absolute time of the first flash and the time (elapsed since that start time) for all flashes you've seen. You should also try to keep track of which flashes are primary and stick to timing the same type of flash.

The second method involves using a tape-recorder, on which you record a radio time signal (beep-beep-beep, yeah). Each time you see a flash, you shout or make some (preferrably short) noise. Afterwards you analyze the tape to write down the times of all flashes. I've never used this method, so if someone wants to suggest improvements over what I've written about it, please go ahead.

An example will clarify things, I hope. This is the format I would like the observations in :

   94    4  Year and month (2I5)
   38.9911  -77.0342      273.  lat, long, hgt (3F10.0)
   14    8   21 25.9   11  Start day,hr,mn,sec  nbr timings
  0.00  0
 18.91  1
 37.35  2
 56.34  3
 74.44  4
 89.01  5
 99.00  6
117.15  7
136.50  8
155.31  10
175.19  12
1 21231U 91029  B 94104.48615656  .00000071  00000-0  57773-4 0  6852
2 21231  82.9474 253.5211 0037446  99.0698 261.4696 13.74759322150327

On the first line year and month are given (April 1994). The second line contains latitude (in decimal degrees, north is positive), and the longitude (in decimal degrees, east is positive) and your height (in meter). The third line contains : day of month (14), hours (8 h UT), minutes (21) and seconds (25.9) of the first flash timed, and also the number of timings you observed (11). The next 11 lines are then the times (seconds since the first time) and index of each of your timings. The index of your timings can be used to indicate if you missed certain flashes (like number 9 and 11 in our fictitious example), if you missed certain flashes (like number 9 and 11 in our fictitious example), or if the flash pattern changed. The last two lines are a set of two line orbital elements of the satellite you observed with epoch around the time of your obs. This is optional, but please do make sure that the name of the satellite is indicated.

The accuracy of all times should be as good as possible, i.e. around 0.2 seconds. Esp. the start time is of paramount importance! If you use the above example as a template for your future observations, that would help the automatization of the analysis a lot. Thanks.

Which passes ?

The aim of the measurements is to determine the direction of the rotation axis reliably and follow its evolution over a longer period. This means that we should try to view the satellite under as diverse geometrical conditions as possible. Mathematically speaking this means that the direction of the bisectrix between the vectors satellite-sun and satellite-obs should be as different as possible from one pass to another. Practically speaking this means that you should try to see as many different passes as possible, e.g. don't always observe the south-to-east pass, but try to observe the west-to-north pass as well. Another example : try to observe not just the zenithal pass, but also the pass that is only 40 degrees above the horizon. This approach will maximize the accuracy obtained during analysis.

It is also useful to follow the satellite over as long a path on the sky as possible, to maximize the chance of seeing a variation in the synodic effect or even of something 'weird'.

Making coordinated simultaneous observations is encouraged, but it is suggested to do this locally, i.e. I won't coordinate those efforts since that seems very impractical. I myself will try to observe simultaneous with some people from Belgium, when weather allows (and that's a very strict condition). I know Mike McCants and Paul Maley will try to do something in Texas. Walter Nissen (in Ohio) and Mark Haun (in the state of Washington) have offered their assistance in the project but lack as yet a 'local' observer to coordinate simultaneous timings with. If you live within a radius of about 1000 km of these people, please contact me, I will put you in contact.

If simultaneous observations seem impossible, you should just try to observe the object as frequently as possible, and send your observations immediately to me ( (don't wait till the end of the month, please).

Which satellites?

95-  2 D, Tsikada rocket
94- 61 B, C1 Kosmos 2292
94- 45 B, C1 Kosmos 2285
94- 41 B, C1 Nadezhda 4
94- 24 B, C1 Kosmos 2279
93- 59 B, J1 Zenit Kosmos 2263
92- 73 B, C1 Kosmos 2218
92- 73 B, C1 Kosmos 2218
92- 64 A, Freja
92- 38 B, SAMPEX Scout rocket
92-  8 B, C1 Kosmos 2180
90- 36 B, C1 Kosmos 2074
90- 23 B, C1 Kosmos 2061
87- 74 G, F2 Tsyklon Kosmos 1875-1880
86- 37 B, C1 Kosmos 1745
85- 41 B, C1 Kosmos 1655
84-109 B, C1 Kosmos 1605
83- 69 J, C1 Kosmos 1473-1480
82- 40 J, C1 Kosmos 1357-1364
70-106 B, Delta/CEP 1 NOAA 1

Most of these satellites have flash periods between 5 and 15 seconds and not-too-difficult to observe, with the exception of Freja which can be quite faint and irregular.

Freja is the only payload in the list. The theory I developed assumes the object to be an end-over-end tumbling cylinder, with flashes coming off of the sides of the cylinder. This is probably an OK assumption for all rockets, but not for payloads. Freja's flashes are probably caused by booms perpendicular to the spin axis, which is completely analogous with the rockets. Since I can also obtain the direction of the rotation axis from my colleagues at MPE and since we *know* the rotation period of Freja to be 6.0 seconds, Freja is the ideal 'control object'.

Quite a few of these objects are 'top-priority' objects in the program of the BWGS, usually because they have been observed to accelerate. Accelerating objects are not necessarily in an end-over-end tumble, fact which may show itself by flashes that are not 'symmetrical', i.e. the time difference between flashes 0 and 1 is not equal to the time difference between flash 1 and 2, but *is* equal to that time difference between flash 2 and 3. There could be two flash periods, following one another. If you notice such behaviour, please tell me, since it may indicate that the assumptions I made are not satisfied for these satellites.

I am open to suggestions for other satellites that you think should be included in this list.


I intend to keep you all updated about changes in the program, first results, etc... through seesat-L (unless the majority of subscribers would rather be spared information about this project) and Flash.

It may all seem more complicated than it really is. I would like to invite everyone to make observations for this project, experience is not a must. It is your chance to contribute to a new field that is scientifically valuable.

Bart De Pontieu

next up previous contents
Next: A closer look Up: Flash 92 Previous: Observations of Molniya